What's Happening?
Citigroup is hiring an Artificial Intelligence (AI) Offensive Security Analyst in Jersey City to enhance the security of its AI systems and leverage AI-powered tools for security testing. This operational role within the Chief Information Security Office
(CISO)'s Offensive Security and Vulnerability Management (OSVM) team involves conducting security assessments against AI systems and utilizing AI tools to improve the speed and effectiveness of security testing. The analyst will be tasked with building tools from scratch, understanding the limits of AI in security, and operating in an environment where established playbooks are still emerging. The position addresses challenges such as AI-driven vulnerability management, securing AI agents that could behave like insider threats, and defining security architecture for the AI era within a global bank that has deployed AI tools to over 180,000 employees.
Why It's Important?
This recruitment highlights Citigroup's proactive approach to addressing the complex security challenges posed by the widespread adoption of AI within its operations. As AI systems become more integrated into critical financial processes, ensuring their security against novel threats like prompt injection, indirect injection, and sensitive data extraction is paramount. The role signifies a recognition that traditional security testing methodologies may not be sufficient for AI, necessitating specialized expertise and AI-powered tools to identify and mitigate vulnerabilities. By focusing on offensive security for AI, Citigroup aims to stay ahead of potential adversaries who might use AI to exploit weaknesses. This initiative is crucial for maintaining the integrity and trustworthiness of financial services in an increasingly AI-driven landscape, protecting client data, and safeguarding the bank's reputation and assets from sophisticated cyber threats.
What's Next?
The AI Offensive Security Analyst will focus on several key areas: testing AI applications and agents, including GenAI applications and LLM-backed products, for specific AI-related attack classes; using AI to enhance security testing by enabling broader coverage, faster reconnaissance, and automated validation of vulnerabilities; and embedding security into the development lifecycle of AI systems. The team aims to build pipelines for sustainable AI-driven vulnerability management, define security architecture for governing agentic AI, and design containment architectures and runtime controls for safely operating AI agents at enterprise scale. This indicates a continuous effort to develop and implement cutting-edge security measures and frameworks tailored specifically for AI technologies within Citigroup.
Beyond the Headlines
Citigroup's investment in AI offensive security reflects a broader industry trend where the rapid deployment of AI necessitates a parallel evolution in cybersecurity strategies. The challenges of securing AI systems, particularly agentic AI that can adapt and potentially act autonomously, introduce new ethical and regulatory considerations. The role's emphasis on 'problems that don’t have playbooks yet' underscores the nascent nature of AI security and the need for innovative, research-driven approaches. This development could lead to the establishment of new industry standards and best practices for AI security, influencing how other financial institutions and highly regulated sectors approach the protection of their AI assets. It also highlights the growing demand for cybersecurity professionals with a deep understanding of AI, signaling a significant shift in the skill sets required for future security roles.














